Nondestructive Analysis of Fluoroquinolone Drugs by Colorimetric Method Using a Smartphone

A method for nondestructive analysis of pharmaceuticals for the content of fluoroquinolones through diffuse reflection of infrared radiation using a smartphone and a device printed on a 3D printer has been proposed. It has been established that the diffuse reflection of infrared radiation with a wavelength of 850 nm from tablets containing active substances of the fluoroquinolone class can be captured using a smartphone camera. The blister packaging and the tablet coating do not hinder the passage of radiation, as confirmed by comparing the results of colorimetric determination of fluoroquinolones in drug samples both in packaging and without it, as well as on a split tablet. A correlation between the analytical signal and the concentration of the active substance is observed regardless of the variant of investigation. The possibility of using chemometric methods is demonstrated, which allows for reduced time of analysis and visualization of the obtained data. The data array was processed using principal component analysis (PCA), hierarchical cluster analysis (HCA), partial least squares regression (PLS), and the least squares method with the use of PhotoMetrix PRO^® software. Using these algorithms, the identification of drugs by their manufacturer was conducted. Colorimetric signals from tablets of the same manufacturer form distinct clusters on diagrams constructed using the HCA algorithm. Data obtained using PCA indicate the placement of signals from tablets of different manufacturers in separate quadrants, allowing for the identification of the pharmaceutical company and the country of manufacture. For example, the drug Ciprofloxacin with varying concentrations of the active substance produced in Russia is located in quadrants 2 and 4, while products manufactured in India are found in quadrants 1 and 3. The use of chemometric analysis methods for determining active substances is discussed in detail. The relative standard deviation of the analysis results did not exceed 0.07.